​
Overview

Bitscale integrates multiple AI-powered agents (BitAgent, Web Search, LLMs) that rely on well-structured prompts to deliver accurate, high-quality outputs. Writing effective prompts ensures you get actionable insights, clean outputs, and scalable workflows for outreach and research. This guide covers how to structure prompts, when to provide context, and how to leverage Bit Prompter, a built-in helper for first-time prompt creation.

​
Structuring a Good Prompt

A strong prompt typically has three sections:
  1. General Directions
    • Start with a concise objective in ~20–30 words.
    • Clearly state what the AI should achieve.
    • Example: “Summarize the website content to highlight product features and flag mentions of compliance.”
  2. Structured Steps
    • Provide explicit instructions on how the AI should process the input.
    • Define what to extract, classify, or generate.
    • Tag important fields like first_name, company_name, etc.
  3. Context Placement
    • For short inputs (e.g., names, titles), include them directly in the prompt.
    • For large inputs (e.g., a full website audit), provide context only after giving directions and steps.
    • This ensures the AI knows what to do before parsing long text.

​
Output Formatting

Unlike many platforms where you must specify output formatting inside the prompt, Bitscale provides a dedicated section for output formatting.
  • Define structured outputs (columns, JSON, bullet points) in the formatting panel.
  • Keep the prompt focused on logic and context, not formatting.

​
Using Bit Prompter

If you’re new to prompt writing, Bitscale provides Bit Prompter, a tool that converts your raw idea into a structured prompt. Example:
  • Input: “Check if a company has SOC2 compliance using their website.”
  • Bit Prompter Output:
    • Objective: Verify SOC2 compliance
    • Directions: Parse website content to detect compliance mentions
    • Steps: Look for keywords like “SOC2,” “audit,” or “certified”
    • Inputs: Website dump (if not already included by BitAgent)
This ensures your prompts follow best practices without requiring deep LLM experience.

​
Best Practices

  • Always begin with a clear objective.
  • Break instructions into steps to reduce ambiguity.
  • Place large context (like long text dumps) after directions.
  • Use output formatting options in Bitscale instead of inline prompt instructions.
  • For consistency, reuse and iterate on prompts that perform well in your workflows.

​
Example 1: Outreach for a New Product Launch

Prompt: 

Write a 100 word email introducing an AI-powered project management tool, 'TaskFlow,' to small business owners. Highlight its unique features: automated task prioritization, deadline reminders, and integration with Google Workspace and Slack.

Structure:

1. Strong opening tailored to small business needs.

2. Brief description of the top 3 features.

3. Invitation to book a demo.

4. Closing with a call-to-action for a free trial.
Why This Works:
  • Targets specific pain points of small business owners.
  • Clearly outlines benefits, increasing the likelihood of engagement.

​
Example 2: Cold Outreach for a Service Company

Prompt: 
Compose a 30 word LinkedIn message for marketing managers in mid-sized tech companies, introducing a content strategy service focused on SEO and lead generation.

Structure:

1. Personalized opening acknowledging their role.

2. Persuasive paragraph about aligning services with their goals.

3. Closing with an invitation to connect or schedule a 15-minute chat.
Why This Works:
  • Creates relevance by referencing the recipient’s role.
  • Offers a clear and actionable next step.

​
Example 3: Follow-Up After No Response

Prompt: 
Write a friendly follow-up email for a PPC advertising company reaching out to B2B SaaS companies.

Structure:

1. Polite reminder referencing the initial email.

2. Highlight key benefits (e.g., conversion optimization).

3. Call-to-action suggesting a brief call with flexible scheduling.

Output Format:

Just Return the mail body with no accompanying texts, salutations, signatures etc.
Why This Works:
  • Maintains a professional yet approachable tone.
  • Reiterates value without pressuring the recipient.

​
Example 4: Outreach Targeting Event Attendance

Prompt: 
Write a LinkedIn message aimed at inviting a marketing director to attend an industry event hosted by our company, a leader in digital marketing analytics. Information of the Event is available in Event\_Info

Structure:

1. Personalized opening that references their interest in digital marketing or data analytics.

2. Brief overview of the event’s key focus and benefits of attending.

3. Friendly closing statement inviting them to RSVP, with a contact link for any questions.

Output Format:

Just Return the message body with no accompanying texts, salutations, signatures etc.

Event_Info = @event
the @ in the above prompt is how we create column instances on Bitscale
Why This Works:
  • Highlights specific value relevant to the recipient’s professional interests.
  • Builds excitement around the event.